Factors associated with measured differences between fourth and fifth phase diastolic blood pressure

Data from the Munich Blood Pressure Study I were used to investigate the distributions of fourth (D4) and fifth (D5) phase diastolic blood pressures in a population and to identify factors related to their difference. Muffling (in contrast to change of amplitude) of sounds was taken as the criterion for fourth phase Korotkoff sounds. D4 could not be detected (ie D4 = D5) in 33.7% of the 1032 men and in 42.1% of the 1163 women. The difference D4-D5 (mean +/- SEM) was 3.1 +/- 0.1 mmHg for male and 2.4 +/- 0.1 mmHg for female participants. Only 28.0% of D4-D5 differences were greater than 4 mmHg in men and 20.2% in women. D4-D5 varied considerably between the three measurements of each examination both related to participants and observers. Participant characteristics associated with greater D4-D5 were higher SBP, higher D4, and lower D5. Smoking was more common in men with greater D4-D5. Inter-observer variability was very strong. Non-detection of D4 ranged from 78.8% to 10.2% between observers. To assess the relative importance of participant and observer influence on the magnitude of D4-D5, we fitted a polychotomous logistic regression model. In this model, participant characteristics had only a weak effect on measured D4-D5. This was outweighted by potential observer effects several-fold stronger than the strongest participant effect. We conclude that for reliability reasons D5 should be given preference over D4 when measuring diastolic blood pressure in adults, whether for clinical or for epidemiological purposes.